The specificity of primer-based amplification reactions, such as the polymerase chain reaction (PCR), largely depends on the specificity of primer hybridization with a DNA template. Under the elevated temperatures used in a typical amplification reaction, the primers ideally hybridize only to the intended target sequence and form primer extension products to produce the complement of the target sequence. However, amplification reaction mixtures are typically assembled at room temperature, well below the temperature needed to insure primer hybridization specificity. Under lower temperature conditions, the primers may bind non-specifically to other partially complementary nucleic acid sequences or to other primers and initiate the synthesis of undesired extension products, which can be amplified along with the target sequence. Amplification of non-specific primer extension products can compete with amplification of the desired target sequences and can significantly decrease the efficiency of the amplification of the desired sequence. Non-specific amplification can also give rise in certain assays to a false positive result.
One frequently observed type of non-specific amplification product in PCR is a template independent artifact of the amplification reaction known as “primer dimers”. Primer dimers are double-stranded fragments whose length typically is close to the sum of the two primer lengths and are amplified when one primer is extended over another primer. The resulting duplex forms an undesired template which, because of its short length, is amplified efficiently.
Non-specific amplification can be reduced by reducing the formation of primer extension products (e.g., primer dimers) prior to the start of the reaction. In one method, referred to as a “hot-start” protocol, one or more critical reagents are withheld from the reaction mixture until the temperature is raised sufficiently to provide the necessary hybridization specificity. In this manner, the reaction mixture cannot support primer extension at lower temperatures. Manual hot-start methods, in which the reaction tubes are opened after the initial high temperature incubation step and the missing reagents are added, are labor intensive and increase the risk of contamination of the reaction mixture.
Alternatively, a heat sensitive material, such as wax, can be used to separate or sequester reaction components, as described in U.S. Pat. No. 5,411,876, and Chou et al., 1992, Nucl. Acids Res. 20(7):1717-1723. In these methods, a high temperature pre-reaction incubation melts the heat sensitive material, thereby allowing the reagents to mix.
Another method of reducing the formation of primer extension products prior to the start of PCR relies on the heat-reversible inactivation of the DNA polymerase. U.S. Pat. Nos. 5,773,258 and 5,677,152, both incorporated herein by reference, describe DNA polymerases reversibly inactivated by the covalent attachment of a modifier group. Incubation of the inactivated DNA polymerase at high temperature results in cleavage of the modifier-enzyme bond, thereby releasing an active form of the enzyme. Non-covalent reversible inhibition of a DNA polymerase by DNA polymerase-specific antibodies is described in U.S. Pat. No. 5,338,671, incorporated herein by reference.
One objective of the present invention can be used, for example, to address the problem of carry-over cross contamination which is a significant concern in amplification reactions, especially PCR wherein a large number of copies of the amplified product are produced. In the prior art, attempts have been made to solve this problem in a number of ways. For example, direct UV irradiation can effectively remove contaminating DNA (Rys & Persing, 1993, J Clin Microbiol. 31(9):2356-60 and Sarkar & Sommer, 1990 Nature. 343(6253):27) but the irradiation of the PCR reagents must take place before addition of polymerase, primers, and template DNA. Furthermore, this approach may be inefficient because the large numbers of mononucleotides present in the reaction will absorb much of the UV light. An alternative, the “UNG method”, incorporates dUTP into the amplified fragments to alter the composition of the product so that it is different from native, naturally occurring DNA (Longo et al. 1990, Gene, 93(1): 125-128). The enzyme Uracil-N-Glycosylase (UNG) is added together with the other components of the PCR mixture. The UNG enzyme will cleave the uracil base from DNA strands of contaminating amplicons before amplification, and render all such products unable to act as a template for new DNA synthesis without affecting the sample DNA. The UNG enzyme is then heat-inactivated and PCR is then carried out. The requirement for dUTP and the UNG enzyme adds significantly to the cost of performing PCR.
Another objective of the present invention is to provide PCR assays in which a hot-start reaction is achieved through a coupled reaction sequence with a thermostable RNase H.
Ribonuclease Enzymes
Ribonucleases (RNases) are enzymes that catalyze the hydrolysis of RNA into smaller components. The enzymes are present internally; in bodily fluids; on the surface of skin; and on the surface of many objects, including untreated laboratory glasswear. Double-stranded RNases are present in nearly all intracellular environments and cleave RNA-containing, double-stranded constructs. Single-stranded RNases are ubiquitous in extracellular environments, and are therefore extremely stable in order to function under a wide range of conditions.
The RNases H are a conserved family of ribonucleases which are present in all organisms examined to date. There are two primary classes of RNase H: RNase H1 and RNase H2. Retroviral RNase H enzymes are similar to the prokaryotic RNase H1. All of these enzymes share the characteristic that they are able to cleave the RNA component of an RNA:DNA heteroduplex. The human and mouse RNase H1 genes are 78% identical at the amino acid level (Cerritelli, et al., (1998) Genomics, 53, 300-307). In prokaryotes, the genes are named rnha (RNase H1) and rnhb (RNase H2). A third family of prokaryotic RNases has been proposed, rnhc (RNase H3) (Ohtani, et al. (1999) Biochemistry, 38, 605-618).
Evolutionarily, “ancient” organisms (archaeal species) in some cases appear to have only a single RNase H enzyme which is most closely related to the modern RNase H2 enzymes (prokaryotic) (Ohtani, et al., J Biosci Bioeng, 88, 12-19). Exceptions do exist, and the archaeal Halobacterium has an RNase H1 ortholog (Ohtani, et al., (2004) Biochem J, 381, 795-802). An RNase H1 gene has also been identified in Thermus thermophilus (Itaya, et al., (1991) Nucleic Acids Res, 19, 4443-4449). RNase H2 enzymes appear to be present in all living organisms. Although all classes of RNase H enzymes hydrolyze the RNA component of an RNA:DNA heteroduplex, the substrate and co-factor requirements are different. For example, the Type II enzymes utilize Mg++, Mn++, Co++ (and sometimes Ni++) as cofactor, while the Type I enzymes require Mg++ and can be inhibited by Mn++ ions. The reaction products are the same for both classes of enzymes: the cleaved products have a 3′-OH and 5′-phosphate (See FIG. 1). RNase III class enzymes which cleave RNA:RNA duplexes (e.g., Dicer, Ago2, Drosha) result in similar products and contain a nuclease domain with similarity to RNase H. Most other ribonucleases, and in particular single stranded ribonucleases, result in a cyclic 2′,3′-phosphate and 5′-OH products (see FIG. 2).
Type I RNase H
E. coli RNase H1 has been extensively characterized. A large amount of work on this enzyme has been carried out, focusing on characterization of substrate requirements as it impacts antisense oligonucleotide design; this has included studies on both the E. coli RNase H1 (see Crooke, et al., (1995) Biochem J, 312 (Pt 2), 599-608; Lima, et al., (1997) J Biol Chem, 272, 27513-27516; Lima, et al., (1997) Biochemistry, 36, 390-398; Lima, et al., (1997) J Biol Chem, 272, 18191-18199; Lima, et al., (2007) Mol Pharmacol, 71, 83-91; Lima, et al., (2007) Mol Pharmacol, 71, 73-82; Lima, et al., (2003) J Biol Chem, 278, 14906-14912; Lima, et al., (2003) J Biol Chem, 278, 49860-49867) and the Human RNase H1 (see Wu, et al., (1998) Antisense Nucleic Acid Drug Dev, 8, 53-61; Wu, et al., (1999) J Biol Chem, 274, 28270-28278; Wu, et al., (2001) J Biol Chem, 276, 23547-23553). In tissue culture, overexpression of human RNase H1 increases potency of antisense oligos (ASOs) while knockdown of RNase H1 using either siRNAs or ASOs decreases potency of antisense oligonucleotides.
Type I RNase H requires multiple RNA bases in the substrate for full activity. A DNA/RNA/DNA oligonucleotide (hybridized to a DNA oligonucleotide) with only 1 or 2 RNA bases is inactive. With E. coli RNase H1 substrates with three consecutive RNA bases show weak activity. Full activity was observed with a stretch of four RNA bases (Hogrefe, et al., (1990) J Biol Chem, 265, 5561-5566). An RNase H1 was cloned from Thermus thermophilus in 1991 which has only 56% amino acid identity with the E. coli enzyme but which has similar catalytic properties (Itaya, et al., (1991) Nucleic Acids Res, 19, 4443-4449). This enzyme was stable at 65° C. but rapidly lost activity when heated to 80° C.
The human RNase H1 gene (Type I RNase H) was cloned in 1998 (Genomics, 53, 300-307 and Antisense Nucleic Acid Drug Dev, 8, 53-61). This enzyme requires a 5 base RNA stretch in DNA/RNA/DNA chimeras for cleavage to occur. Maximal activity was observed in 1 mM Mg++ buffer at neutral pH and Mn++ ions were inhibitory (J Biol Chem, 274, 28270-28278). Cleavage was not observed when 2′-modified nucleosides (such as 2′-OMe, 2′-F, etc.) were substituted for RNA.
Three amino acids (Asp-10, Glu-48, and Asp-70) make up the catalytic site of E. coli RNase H1 which resides in the highly conserved carboxy-terminal domain of the protein (Katayanagi, et al., (1990) Nature, 347, 306-309); this domain has been evaluated by both site directed mutagenesis and crystal structure determination. The same amino acids are involved in coordination of the divalent ion cofactor.
Interestingly, 2′-modification of the substrate duplex alters the geometry of the helix and can adversely affect activity of RNase H1. 2′-O-(2-methoxyl)ethyl (MOE) modifications flanking the RNA segment reduce cleavage rates, presumably due to alterations in the sugar conformation and helical geometry. Locked nucleic acid (LNA) bases perturb helical geometry to a greater degree and impacted enzyme activity to a greater extent (Mol Pharmacol, 71, 83-91 and Mol Pharmacol, 71, 73-82). Damha (McGill University) has studied the effects of 2′-F modified nucleosides (2′-deoxy-2′-fluoro-b-D-ribose) when present in the substrate duplex and finds that this group cannot be cleaved by RNase H1 (Yazbeck, et al., (2002) Nucleic Acids Res, 30, 3015-3025). Formulas A and B illustrate the two different mechanisms that have been proposed for RNase H1 cleavage, both of which require participation of the 2′OH group.

Damha's studies are consistent with the known active site of the enzyme, wherein the reaction mechanism involves the 2′-OH group. The enzyme active site resides within a cluster of lysine residues which presumably contribute to electrostatic binding of the duplex. Interaction between the binding surface and negatively charged phosphate backbone is believed to occur along the minor grove of the RNA:DNA heteroduplex (Nakamura, et al., (1991) Proc Natl Acad Sci USA, 88, 11535-11539); changes in structure that affect the minor groove should therefore affect interactions between the substrate and the active site. For example, the minor groove width is 7.5 Å in a B-form DNA:DNA duplex, is 11 Å in a pure A-form RNA:RNA duplex, and is 8.5 Å in the hybrid A-form duplex of an RNA:DNA duplex (Fedoroff et al., (1993) J Mol Biol, 233, 509-523). 2′-modifications protrude into the minor groove, which may account for some of the behavior of these groups in reducing or eliminating activity of modified substrates for cleavage by RNase H1. Even a 2′-F nucleoside, which is the most “conservative” RNA analog with respect to changing chemical structure, adversely affects activity.
Type II RNase H
The human Type II RNase H was first purified and characterized by Eder and Walder in 1991 (Eder, et al., (1991) J Biol Chem, 266, 6472-6479). This enzyme was initially designated human RNase H1 because it had the characteristic divalent metal ion dependence of what was then known as Class I RNases H. In the current nomenclature, it is a Type II RNase H enzyme. Unlike the Type I enzymes which are active in Mg++ but inhibited by Mn++ ions, the Type II enzymes are active with a wide variety of divalent cations. Optimal activity of human Type II RNase H is observed with 10 mM Mg++, 5 mM Co++, or 0.5 mM Mn++.
Importantly, the substrate specificity of the Type II RNase H (hereafter referred to as RNase H2) is different from RNase H1. In particular, this enzyme can cleave a single ribonucleotide embedded within a DNA sequence (in duplex form) (Eder, et al., (1993) Biochimie, 75, 123-126). Interestingly, cleavage occurs on the 5′ side of the RNA residue (See FIG. 3). See a recent review by Kanaya for a summary of prokaryotic RNase H2 enzymes (Kanaya (2001) Methods Enzymol, 341, 377-394).
The E. coli RNase H2 gene has been cloned (Itaya, M. (1990) Proc Natl Acad Sci USA, 87, 8587-8591) and characterized (Ohtani, et al., (2000) J Biochem (Tokyo), 127, 895-899). Like the human enzyme, the E. coli enzyme functions with Mn++ ions and is actually more active with manganese than magnesium.
RNase H2 genes have been cloned and the enzymes characterized from a variety of eukaryotic and prokaryotic sources. The RNase H2 from Pyrococcus kodakaraensis (KOD1) has been cloned and studied in detail (Haruki, et al., (1998) J Bacteriol, 180, 6207-6214; Mukaiyama, et al., (2004) Biochemistry, 43, 13859-13866). The RNase H2 from the related organism Pyrococcus furious has also been cloned but has not been as thoroughly characterized (Sato, et al., (2003) Biochem Biophys Res Commun, 309, 247-252).
The RNase H2 from Methanococcus jannaschii was cloned and characterized by Lai (Lai, et al., (2000) Structure, 8, 897-904; Lai et al., (2003) Biochemistry, 42, 785-791). Isothermal titration calorimetry was used to quantitatively measure metal ion binding to the enzyme. They tested binding of Mn++, Mg++, Ca++, and Ba++ and in all cases observed a 1:1 molar binding ratio, suggesting the presence of only a single divalent metal ion cofactor in the enzyme's active site. The association constant for Mn++ was 10-fold higher than for Mg++. Peak enzyme activity was seen at 0.8 mM MnCl2.
Nucleic acid hybridization assays based on cleavage of an RNA-containing probe by RNase H such as the cycling probe reaction (Walder et al., U.S. Pat. No. 5,403,711) have been limited in the past by background cleavage of the oligonucleotide by contaminating single-stranded ribonucleases and by water catalyzed hydrolysis facilitated by Mg2+ and other divalent cations. The effect of single-stranded ribonucleases can be mitigated to a certain degree by inhibitors such as RNasin that block single-stranded ribonucleases but do not interfere with the activity of RNase H.
Single-stranded ribonucleases cleave 3′ of an RNA residue, leaving a cyclic phosphate group at the 2′ and 3′ positions of the ribose (See FIG. 2). The same products are produced by spontaneous water catalyzed hydrolysis. In both cases, the cyclic phosphate can hydrolyze further forming a 3′-monophosphate ester in the enzyme catalyzed reaction, or a mixture of the 3′- and 2′-monophosphate esters through spontaneous hydrolysis. The difference between the cleavage products formed by RNase H (FIG. 1) and those formed by nonspecific cleavage of the probe (FIG. 2) provides a basis for distinguishing between the two pathways. This difference is even more pronounced when comparing cleavage by RNase H2 and single-stranded ribonucleases with substrates having only a single RNA residue. In that case, RNase H2 and single-stranded ribonucleases attack at different positions along the phosphate backbone (See FIG. 3).
RNase H has been used as a cleaving enzyme in cycling probe assays, in PCR assays (Han et al., U.S. Pat. No. 5,763,181; Sagawa et al., U.S. Pat. No. 7,135,291; and Behlke and Walder, U.S. Pat. App. No. 20080068643) and in polynomial amplification reactions (Behlke et al., U.S. Pat. No. 7,112,406). Despite improvements offered by these assays, there remain considerable limitations. The PCR assays utilize a hot-start DNA polymerase which adds substantially to the cost. Moreover, each time an alternative DNA polymerase is required a new hot-start version of the enzyme must be developed. In addition, the utility of these various assays has been limited by undesirable cleavage of the oligonucleotide probe or primer used in the reaction, including water and divalent metal ion catalyzed hydrolysis 3′ to RNA residues, hydrolysis by single-stranded ribonucleases and atypical cleavage reactions catalyzed by Type II RNase H enzymes at positions other than the 5′-phosphate of an RNA residue. The present invention overcomes these limitations and offers further advantages and new assay formats for use of RNase H in biological assays.
The current invention provides novel biological assays that employ RNase H cleavage in relation to nucleic acid amplification, detection, ligation, sequencing, and synthesis. Additionally, the invention provides new assay formats to utilize cleavage by RNase H and novel oligonucleotide substrates for such assays. The compounds, kits, and methods of the present invention provide a convenient and economic means of achieving highly specific primer-based amplification reactions that are substantially free of nonspecific amplification impurities such as primer dimers. The methods and kits of the present invention avoid the need for reversibly inactivated DNA polymerase and DNA ligase enzymes.